[ad_1]
If you’re utilizing the extensively and wildly common AI app ChatGPT, or if you’re planning on utilizing it, take into account that a lawsuit would possibly ultimately be headed your method.
Right here’s the deal.
Suppose you choose to utilize ChatGPT in the midst of crafting or delivering a service of 1 variety or one other. Lots of people are doing so nowadays. The concept is that ChatGPT can bolster your effectivity and make you extra productive. This would possibly imply much less time required to get your job accomplished. It may also imply that you are able to do extra work than you had beforehand achieved, hopefully incomes you some more money too.
For these of you unfamiliar with this hottest and newest AI app, ChatGPT is a headline-grabber that’s extensively recognized for with the ability to produce fluent essays and keep it up interactive dialogues, virtually as if being undertaken by human fingers. An individual enters a written immediate, ChatGPT responds with a couple of sentences or a whole essay, and the ensuing encounter appears eerily as if one other particular person is chatting with you somewhat than an AI utility. That is known as generative AI because it generates textual content or essays in response to text-entered prompts. ChatGPT is made by the agency OpenAI, an organization that has change into the darling of the AI business and garners all method of avid consideration nowadays.
To get extra particulars about how ChatGPT works, see my rationalization at the link here. If you’re within the successor to ChatGPT, coined GPT-4, see the dialogue at the link here.
Generative AI relies on a fancy computational algorithm that has been information educated on textual content from the Web and admittedly can do some fairly spectacular pattern-matching to have the ability to carry out a mathematical mimicry of human wording and pure language. Please notice that ChatGPT isn’t sentient. We don’t have sentient AI. Don’t fall for these zany headlines and social media rantings suggesting in any other case.
With that lowdown, we are able to now get again to the notion of utilizing ChatGPT as a part of your work effort or religious utilization in different methods.
Let’s say that somebody comes ahead and claims that your service or work effort has allegedly brought on them hurt.
They resolve to sue you.
As an apart, none of us needs or relishes that such a declare would possibly come up. You may need been doing the very best that you are able to do. You would possibly really feel totally wronged as to the declare being made. Nonetheless, on this zany lawsuit mania world that we reside in, there may be at all times a viable probability that somebody will contend they have been harmed and search to sue you accordingly. That’s the tough actuality.
The particular person suing you would possibly notice that you just do not have deep pockets, diminishing the chances of getting a lot dough out of you. They will ergo angle towards one thing that does have tons of greenbacks. One thing that might be a juicy goal for the lawsuit. One thing that someway pertains to your service and so they contend contributed straight or not directly to the alleged hurt concerned.
Effectively, naturally, they’d search to sue OpenAI (the maker of ChatGPT) and attempt to embroil them into the lawsuit too.
Sure, certainly, OpenAI has the large bucks. You could have the small bucks. The particular person launching the lawsuit would ardently argue that your efforts have been aided by ChatGPT and due to this fact each you and OpenAI must be on the hook. The posturing can be that have been it not for ChatGPT, presumably, you wouldn’t have been capable of give rise to the hurt that you just purportedly brought on.
I’d guess that you just’ve seen this general state of affairs many instances within the information. An individual leveraging this or that service or product is sued for alleged hurt they’ve brought on. The lawsuit additionally names the maker of the underlying service or product. A cynic would say it’s as a result of the maker of the underlying contrivance has the deeper pockets. Others would possibly say that the maker of the underlying aspect ought to rightfully be partially at fault. It’s a truthful and sq. justification to incorporate the maker within the authorized motion too, they’d fervently proclaim.
Your first thought is likely to be that you would care much less that OpenAI can also be encompassed within the lawsuit.
Your fear is about you. Let OpenAI defend itself. OpenAI undoubtedly has legions of legal professionals and luggage of coinage to rent outdoors attorneys. They’ll just do effective. You, alternatively, barely make your mortgage funds. Having to defend your self in a doubtlessly pricey lawsuit is unnerving and doubtlessly may price you every thing you’ve ever earned.
I’m about to do the reveal right here on this thorny matter, so put together your self.
It is a set off warning.
You won’t have seemed carefully on the licensing phrases related together with your signing up to make use of ChatGPT. Most individuals don’t. They assume that the licensing is the standard legalese that’s impenetrable. Plus, the belief is that there’s nothing in there that can be worthy of explicit consideration. Simply the standard ramblings of arcane authorized stuff.
Effectively, you would possibly need to take into account Part 7a of the present licensing settlement as posted on the OpenAI web site and related to and encompassing your use of ChatGPT:
- “Part 7. Indemnification; Disclaimer of Warranties; Limitations on Legal responsibility: (a) Indemnity. You’ll defend, indemnify, and maintain innocent us, our associates, and our personnel, from and towards any claims, losses, and bills (together with attorneys’ charges) arising from or regarding your use of the Companies, together with your Content material, services or products you develop or supply in reference to the Companies, and your breach of those Phrases or violation of relevant legislation.”
In regular language, this usually means that if OpenAI will get sued for one thing you’ve got accomplished with their providers or merchandise similar to ChatGPT, you’re thought-about by them to be on the hook for “any claims, losses and bills (together with attorneys’ charges)” thereof.
Backside line, you may need to cowl your individual authorized bills plus no matter monetary hit you are taking from the lawsuit, and moreover doubtlessly cowl the authorized bills and associated monetary hit that OpenAI incurs as a result of lawsuit.
That ought to provide you pause for thought.
Plenty of pauses. Plenty of thought.
That is the hidden double whammy that few notice sits silently inside their ongoing use of ChatGPT.
To make clear, by and enormous, many of the on-line providers that you just use are prone to have an identical clause. This isn’t particularly uncommon or distinctive. That is shall we embrace customary. You most likely didn’t notice that was the case. Nor have you ever most likely ever discovered your self really in actuality topic to the indemnification clause.
Good for you.
Fortunate for you.
If you’re utilizing ChatGPT, you ought to pay attention to this potential double publicity. Embody this consider your startup concerns if you’re a budding entrepreneur. I might additionally urge that you need to have already got had an lawyer advise you about your use of ChatGPT, assuming that you’re utilizing the AI app in some capability that could possibly be linked to one thing you would possibly do and probably give rise to a claimed hurt.
I do know what a few of you’re pondering. Because you didn’t know concerning the indemnification clause, you may seemingly escape it by declaring truthfully and sincerely that you just by no means knew it was there. Had you realized it was there, you’ll have acted otherwise. No person can anticipate you to be in your toes about one thing that you just had no information of. That’s plainly the one truthful solution to see issues.
I’ve a notable quote for you.
Addison Mizner, the well-known architect, stated this concerning the legislation: “Ignorance of the legislation excuses no particular person from working towards it.”
Making an attempt to weasel your method out of the matter by pleading ignorance is an extended and bumpy street. Ask your lawyer and see what they are saying about how far an ignorance of the legislation protection can get you. Often, not very far.
In right now’s column, I’ll stroll you thru a few of the intricacies related to indemnification clauses. This is likely to be useful to you as you’re day by day making use of ChatGPT or are contemplating doing so. Please do notice it is a authorized matter of unbelievable depth and I’m solely protecting the floor of what the matter entails. Your greatest guess, as indicated earlier, can be to at all times and immediately be sure you have authorized counsel that may advise you on these sorts of thorny authorized issues.
You is likely to be tempted to attend till a lawsuit comes your solution to then search authorized recommendation. Sorry to say that that is the proverbial mistake of letting the horse out of the barn. You’ll be wiser and safer to line up your geese beforehand. Be prepared for the day that the storm brews. Don’t get caught within the open and struck by bolts of lightning.
Into all of this comes a slew of AI Ethics and AI Regulation concerns.
There are ongoing efforts to imbue Moral AI rules into the event and fielding of AI apps. A rising contingent of involved and erstwhile AI ethicists are attempting to make sure that efforts to plot and undertake AI takes into consideration a view of doing AI For Good and averting AI For Unhealthy. Likewise, there are proposed new AI legal guidelines which can be being bandied round as potential options to maintain AI endeavors from going amok on human rights and the like. For my ongoing and intensive protection of AI Ethics and AI Regulation, see the link here and the link here, simply to call a couple of.
The event and promulgation of Moral AI precepts are being pursued to hopefully forestall society from falling right into a myriad of AI-inducing traps. For my protection of the UN AI Ethics rules as devised and supported by almost 200 international locations through the efforts of UNESCO, see the link here. In an identical vein, new AI legal guidelines are being explored to attempt to preserve AI on a fair keel. One of many newest takes consists of a set of proposed AI Invoice of Rights that the U.S. White Home not too long ago launched to determine human rights in an age of AI, see the link here. It takes a village to maintain AI and AI builders on a rightful path and deter the purposeful or unintentional underhanded efforts that may undercut society.
I’ll be interweaving AI Ethics and AI Regulation associated concerns into this dialogue.
Getting Blindsided From All Sides
I’ve beforehand lined the quite a few banned makes use of of ChatGPT that OpenAI says you aren’t imagined to be doing, see the link here. If you’re certainly making use of ChatGPT in any of these prohibited methods, you’re already in dicey waters. Even in the event you aren’t utilizing ChatGPT in these bitter and dour methods, you may nonetheless be utilizing the commemorated AI app in seemingly absolutely professional methods and change into the topic of a lawsuit by somebody that believes you’ve got brought on them hurt on account of your ChatGPT use (or so they could declare).
Listed below are the banned makes use of in a nutshell and I present descriptions of every in my column protection:
- 1) Nothing Unlawful
- 2) Not Any Youngster Exploitation
- 3) Not Hateful
- 4) No Malware
- 5) No Bodily Hurt
- 6) No Financial Hurt
- 7) No Fraud
- 8) No Grownup Content material
- 9) No Political Campaigning Or Lobbying
- 10) No Privateness Intrusion
- 11) No Unauthorized Apply Of Regulation
- 12) No Unaided Monetary Recommendation
- 13) No Improper Well being Recommendation
- 14) Not For Excessive-Threat Governing
- 15) Different Precautions
I suppose the very best guess is to keep away from the thought-about prohibited makes use of. That’s one smart step amongst others that I’ll point out herein.
In the case of probably getting sued on account of your providers or different efforts, and if these providers or efforts are not directly or straight formed on account of utilizing ChatGPT, these are the circumstances you would possibly regrettably face:
- You get sued. You alone are sued (let’s additionally assume that you’re not directly or straight making use of ChatGPT in some associated method)
- You and OpenAI get sued. You might be sued and OpenAI because the maker of ChatGPT can also be sued too
- Solely OpenAI will get sued. OpenAI because the maker of ChatGPT is sued, however you aren’t sued, after which OpenAI involves you to cowl their lawsuit prices on account of the indemnification clause and an assertion that you just spurred the lawsuit
- Different variations
One huge query can be whether or not you really are utilizing ChatGPT and whether or not you’re doing so in a way that pertains to the providers or work efforts that supposedly have led to the claimed hurt. Should you haven’t been utilizing ChatGPT in any respect, this would appear to not less than unleash you from the OpenAI aspects, although there may be sure to be murkiness tossed into these waters.
When you’ve got been utilizing ChatGPT, the subsequent stage of scrutiny would doubtless be whether or not the utilization in any respect pertains to the providers or work efforts which can be alleged to have brought on the hurt.
In response to the OpenAI web site that signifies the licensing settlement related to their merchandise together with ChatGPT, they point out this concerning the providers that they’re offering to you (assuming you make use of ChatGPT):
- “These Phrases of Use apply whenever you use the providers of OpenAI, L.L.C. or our associates, together with our utility programming interface, software program, instruments, developer providers, information, documentation, and web sites (“Companies”). The Phrases embody our Service Phrases, Sharing & Publication Coverage, Utilization Insurance policies, and different documentation, pointers, or insurance policies we might present in writing. By utilizing our Companies, you agree to those Phrases.”
Moreover, this is similar web site depiction of the content material that you’re producing on account of utilizing ChatGPT:
- “Part 3. Content material: (a) Your Content material. Chances are you’ll present enter to the Companies (‘Enter’), and obtain output generated and returned by the Companies primarily based on the Enter (‘Output’). Enter and Output are collectively ‘Content material.’ As between the events and to the extent permitted by relevant legislation, you personal all Enter. Topic to your compliance with these Phrases, OpenAI hereby assigns to you all its proper, title and curiosity in and to Output. This implies you should use Content material for any objective, together with business functions similar to sale or publication, in the event you adjust to these Phrases. OpenAI might use Content material to offer and preserve the Companies, adjust to relevant legislation, and implement our insurance policies. You might be liable for Content material, together with for guaranteeing that it doesn’t violate any relevant legislation or these Phrases.”
You would possibly discover of curiosity my evaluation of these phrases as a matter of whether or not or not the ChatGPT essays are doubtlessly in violation of Mental Property (IP) legal guidelines, plus doubtlessly a type of plagiarism, see my protection at the link here.
As a reminder of the sooner cited Part 7a encompassing indemnification when you find yourself making use of ChatGPT, right here’s what that portion says:
- “Part 7. Indemnification; Disclaimer of Warranties; Limitations on Legal responsibility: (a) Indemnity. You’ll defend, indemnify, and maintain innocent us, our associates, and our personnel, from and towards any claims, losses, and bills (together with attorneys’ charges) arising from or regarding your use of the Companies, together with your Content material, services or products you develop or supply in reference to the Companies, and your breach of those Phrases or violation of relevant legislation.”
Many individuals are usually not certain of what indemnification as a authorized principle means or implies.
Typically, the notion of indemnification is that you’re agreeing to assist in making complete another person that has incurred a loss or injury, right here’s a useful formalized definition:
- “Recompense for loss, injury, or accidents; restitution or reimbursement. An indemnity contract arises when one particular person takes on the duty to pay for any loss or injury that has been or is likely to be incurred by one other particular person. The appropriate to indemnity and the responsibility to indemnify ordinarily stem from a contractual settlement, which usually protects towards legal responsibility, loss, or injury” (posted in The Free Dictionary, West’s Encyclopedia of American Regulation).
There are normally two events concerned in an indemnification clause, whereby one get together, we’ll name this generically Occasion A (the particular person or entity looking for to be indemnified), desires to have assurances from one other get together, we’ll name this generically Occasion B (the particular person or entity that’s endeavor the indemnification), that if one thing occurs to Occasion A on account of actions by Occasion B then the Occasion B will cowl the lack of Occasion A accordingly.
Think about that you’re Occasion B, whereas OpenAI is Occasion A.
You resolve to signup to make use of ChatGPT which is an AI app being supplied by OpenAI. There may be an indemnification clause within the licensing coverage of ChatGPT. You, as Occasion B, have ostensibly agreed to indemnify Occasion A, OpenAI, upon continuing to utilize ChatGPT.
The considerably obvious occasion consists of a lawsuit towards you and your use of ChatGPT which additionally then consists of OpenAI as a part of the lawsuit. In flip, OpenAI involves you to cowl them concerning the lawsuit. They’ve invoked the indemnification clause.
One other avenue is that suppose OpenAI will get sued for one thing about ChatGPT, and for which it’s claimed that your use of ChatGPT is integral to the contended hurt that’s the foundation for the lawsuit. OpenAI, as Occasion A, can search to invoke the indemnification clause and are available to you for protection of the related prices and whatnot.
The American Bar Affiliation (ABA), explains indemnification general through this depiction:
- “Indemnification is the follow of guaranteeing a 3rd get together declare towards your counterparty. Think about that you’ve a contract with a staffing company to provide momentary workers working in your property, and in the midst of their assigned duties, a type of temps causes a 3rd get together to be injured. The injured third get together sues you and the staffing company and secures compensation for private accidents. Each you and the seller have monetary legal responsibility in some proportion in consequence, nonetheless your contract required the staffing company to indemnify you for any third get together claims that arose within the efficiency of the contract. Because of this the staffing company will take over the complete legal responsibility for the damages—they’ve indemnified you for the loss” (an ABA posted article entitled “Negotiating Indemnity” by Taylor Brown, Might 5, 2017).
A state of affairs of how this might doubtlessly apply within the use case of ChatGPT is sketched in an insightful analysis article by Roee Sarel entitled “Restraining ChatGPT” and posted as a Working Paper that’s forthcoming to be printed within the UC Regulation SF Journal (previously Hastings Regulation Journal).
The analysis article posits this lifelike state of affairs. Jack, a lawyer, mistakenly depends on the output of ChatGPT for a authorized matter and consists of the textual content in a authorized briefing filed with the court docket.
I’ve mentioned in my columns at size the priority that persons are sure to be lulled into assuming that ChatGPT output is appropriate. ChatGPT interactions and output is normally conveyed with a tone and sense of confidence that belies the reality that ChatGPT outputs can comprise errors, biases, falsehoods, and so-called AI hallucinations. Should you get snug utilizing ChatGPT and perchance haven’t but encountered these maladies, your strident perception that ChatGPT can do no flawed goes method up. Finally, you let down your guard and simply settle for that the ChatGPT outputs are good to go. Positive, you would possibly look on the outputted essays, however if you’re in a rush, you accomplish that in a cursory method.
This could additionally occur to legal professionals. My urging is that legal professionals might want to particularly keep on their guard when utilizing ChatGPT, see my protection at the link here and the link here, simply to call a couple of. A lawyer can discover that utilizing ChatGPT or different generative AI is a giant time saver, permitting the lawyer to get work accomplished expeditiously and impress their purchasers accordingly. The road right now is that lawyer’s utilizing generative AI will outdo legal professionals that aren’t utilizing generative AI. The catch, maybe, can be that you’ll want to be cautious when utilizing ChatGPT and double-check or triple-check no matter outputs are generated.
Let’s get again to Jack, our ChatGPT-using lawyer within the state of affairs of the analysis article.
Upon having filed a doc with the court docket that accommodates ChatGPT outputs that turned out to be defective, the analysis article makes this key level:
- “What’s the hurt of such a mistake? Clearly, Jack himself might endure a reputational loss in case the choose reprimands him for deceptive the court docket, which can even discover its method into the court docket protocol. After all, such hurt is considerably uninteresting, as it’s anyway ruled by the contractual relationship between Jack and ChatGPT by means of the phrases of service. As of February 7, 2023, the phrases of service certainly not solely state that OpenAI (the creators of ChatGPT) bear no legal responsibility however that the customers of the chatbot should indemnify OpenAI for any third-party claims” (ibid).
You possibly can hopefully plainly see that the indemnification clause of OpenAI comes into the image.
Right here’s an extra evaluation contained within the analysis article on that side:
- “Thus, the extra fascinating query is what exactly are these third-party claims. Within the instance of Jack the lawyer, his consumer might resolve so as to add ChatGPT to a malpractice lawsuit, blaming not solely the lawyer for negligence but in addition the AI itself for offering inaccurate outcomes. As Jack’s consumer doesn’t have a contractual relationship with OpenAI, such a lawsuit would doubtless be primarily based on a tort declare that factors to the consumer’s loss because the related hurt” (ibid).
As famous, let’s assume that the consumer of Jack that was doubtlessly harmed in some vogue on account of the defective submitting with the court docket, turns round and sues Jack for authorized malpractice. The consumer may also decide to sue OpenAI. In the event that they sued OpenAI, presumably the AI maker would come to Jack and invoke the indemnification clause, looking for to have Jack cowl the prices and whatnot to defend towards the lawsuit of Jack’s consumer.
Fairly a multitude.
The analysis article asks an as but unsettled query about whether or not it’s even primarily truthful that OpenAI ala ChatGPT is being included in such a lawsuit:
- “Within the instance of Jack the lawyer, the hurt is concentrated together with his consumer and happens instantly, however the direct injurer is the lawyer and never ChatGPT. Therefore, it isn’t apparent that the AI creators owe an obligation of care to the consumer, who is simply an oblique sufferer of misinformation delivered to the lawyer underneath the express contractual situation that the creator isn’t liable.”
It’s all a authorized hornet’s nest.
The kind of hornet’s nest that you just most likely need to attempt to keep away from getting ensnared in (properly, are they any hornet’s nests that we do need to get entangled in, one would possibly ask rhetorically).
Making an attempt To Contend With The Dreadful Predicament
What are you to do about all of this?
First, just be sure you take a look at the licensing settlement for any software program or AI app that you just use, together with all that enjoyable and fascinating on-line stuff that catches your consideration. Is there an indemnification clause? If that’s the case, take into accout what that may foretell.
Second, take into account not utilizing the AI app as one possibility, thus avoiding getting embroiled afterward in a messy authorized affair. Test round to see if a comparable AI app doesn’t have such a provision. All else being equal, doubtless select that AI app over the opposite one. I can although let you know that the chances are that almost all AI apps which can be of a outstanding nature are going to comprise such a clause. You most likely gained’t discover a comparable AI app that lacks it, however I definitely say it’s price wanting round simply in case.
Third, seek the advice of your lawyer about find out how to greatest attempt to legally defend your self if you’re insistent that you should use ChatGPT or another generative AI that accommodates such a clause. Higher to be ready. When a lawsuit comes, the shock is likely to be lessened in the event you’ve accomplished your homework beforehand. You possibly can not less than have been embarking on a path that can allow your lawyer to maybe discover methods to bolster your protection.
Fourth, it’s conceivable that the AI maker of ChatGPT or the maker of no matter different generative AI app that you’re utilizing would possibly decide to not invoke the indemnification clause. There is likely to be smart causes for the AI maker to not accomplish that.
Think about this excerpt from an article entitled “Will On-line Indemnification Agreements Be Enforced?” about whether or not on-line indemnification clauses are thought-about legally enforceable or not:
- “Most customers in the US buy items and providers on-line. And most of them by no means learn the ‘phrases and situations’ which can be usually embedded of their transactions. But, these unread phrases can embody essential obligations concerning future disputes, similar to provisions requiring the patron to indemnify the seller and maintain it innocent. Are such on-line indemnification provisions enforceable? Little case legislation straight addresses this query. To make sure, courts have upheld some on-line indemnification provisions” (article by Julie Cilia in The Girl Advocate, Quantity 21, Quantity 3, Spring 2016).
There’s a slew of the way to attempt to undercut or vacate an indemnification clause, similar to however not restricted to:
- Jurisdictional dispute as to countermanding provisions on the federal versus state ranges
- Shopper safety provisions that may apply
- Probably obscure and imprecise language of the clause
- Improper or legally faulty language of the clause
- Lack of appropriate constructive discover (the settlement is hidden or onerous to search out)
- Lack of mutual manifestation of assent (did each events should have a gathering of the minds)
- Lack of particularly expressed assent (i.e., when the licensing is discovered through a hyperlink or browsewrap, versus using clickwrap the place a person should click on earlier than they’ll proceed into utilizing the app, or scroll wrap the place you’ll want to scroll after which click on to affirm)
- Overdose of adhesion similar to a take-it-or-leave-it and no negotiating allowed
- Provision wasn’t sufficiently triggered or was improperly invoked
- Unconscionable as to “clean cheque” or an uncapped onerous monetary burden
- Negligence or failure on a part of the service supplier
- And so forth.
There’s a well-known quote that legal professionals are likely to know by coronary heart that generally the courts will take into account such issues as seemingly unconscionable and so outrageously unfair as to ardently shock the judicial conscience (see for instance the authorized case of Tune fi, Inc. v. Google Inc. & YouTube, LLC). Thus, there’s a probability that you just would possibly discover sympathy from the court docket and be relieved of the burden of an indemnification clause, although this isn’t in any respect an ironclad consequence and as a substitute a veritable roll of the cube.
Conclusion
The thrill about ChatGPT has spurred many enterprising entrepreneurs to incorporate of their serving choices a modicum of utilization of ChatGPT. The goal is to spur their providers and garner extra clients. As well as, there’s a stable probability that they’ll present the providers at a decrease price or on a speedier foundation as a consequence of utilizing generative AI similar to ChatGPT.
Utilizing ChatGPT additionally gives a advertising and marketing angle. You possibly can tout that you’re utilizing the red-hot and hottest generative AI of right now. It is a bandwagon that gleans eyeballs and media curiosity.
Of their haste to leap onto that bandwagon, few have a tendency to provide due consideration to the licensing agreements underlying numerous generative AI apps and are sadly exposing themselves to monetary and reputational dangers. They won’t realize it on the time. The dismay and shock will come afterward, assuming that somebody seeks to sue them and the AI maker for an alleged flawed.
Just be sure you have gotten your authorized concerns found out, upfront and on an ongoing foundation too. There are many authorized landmines ready for you down the entrepreneurial path, together with when utilizing generative AI. Sound authorized recommendation can assist in dealing with these authorized exposures.
A closing comment for now on this matter.
There may be an acclaimed authorized joke by the one-liner comic Steven Wright: “I busted a mirror and bought seven years of dangerous luck, however my lawyer thinks they’ll get me 5.”
It’s onerous to stay lighthearted whenever you get tangled in a authorized dispute, however in the event you preserve your eyes open and take suggested authorized precautions, hopefully, you’ll have much less trauma if the mirror does crack or break.
[ad_2]
Source link