ChatGPT - chatbot

Transparency, accountability 🙈

ChatGPT is seen to have a number of important transparency and accountability limitations, ranging from so-called "black box" decision-making to OpenAI's opaque corporate governance.

• Corporate governance. OpenAI is seen as highly opaque about many aspects of its corporate governance, including its funding, financial performance and supply chain management - for ChatGPT and its other products and services.

• Black box decision-making. ChatGPT is a de facto "black box" system, making it difficult to understand how it arrives at specific outputs. Access to its data, model and algorithm are jealously guarded.

• User information. Research studies show that ChatGPT is amongst the least open large language models, with little public information available about how it works or what its impacts are.

• Privacy. Italy's privacy regulator pointed out OpenAI appears to have been made deliberately difficult to find privacy-related information, make complaints and request the removal of personal information.

• Data sources. OpenAI refuses to divulge information about the sources it uses to train ChatGPT and other systems, possibly as it may have appropriated copyrighted data without acknowledgement or permission. 

• Environmental impacts. OpenAI provides no information about the amount of energy and water are required to train and run its systems, including ChatGPT, and the downstream impacts of this consumption.

• Complaints and appeals. Many users have reported a noticeable decline in the quality of responses from ChatGPT after significant updates, leading to high volumes of complaints about the difficulties associated with both using the service and OpenAI's apparent reluctant to acknowledge and address these issues.

Risks, harms 🛑

ChatGPT has been criticised for a wide range of actual and potential harms, including: 

• Anthropomorphism. ChatGPT's ability to generate plausible-looking, human-like information and advice runs the risk of convincing users that it is somehow human and/or 'sentient'. 

• Bias and discrimination. Despite built-in safeguards, researchers and others have demonstrated that ChatGPT produces outputs that are explicitly and implicitly biased in multiple ways.

• Accuracy and reliability. Including 'hallucinations', or plausible-sounding inaccurate information such as fake references or sources.

• Copyright. OpenAI has been widely accused of using copyright protected information without permission to train its language models and ChatGPT. 

• Misinformation and disinformation. ChatGPT has been shown to produce large volumes of convincing misinformation and disinformation.

• Unsafe and offensive content. ChatGPT has safeguards to minimise it being tricked into making errors and to stop requests generating inappropriate, abusive, or offensive content, including hate speech. However, it appears some of these can be overridden relatively easily.

• Confidentiality. Unless they expressly opt-out, OpenAI uses user information and data inputted to ChatGPT to train its models, thereby potentially or actually violating their personal confidentality or the confidentiality of their employers.

• Privacy. Concerns exist about ChatGPT's approach to and impact on personal privacy. 

• Reputational damage. The use and misuse of ChatGPT can harm the reputation of individuals and organisations it generates information about, in some instances resulting in threats to sue OpenAI for defamation. 

• Security. ChatGPT has been shown to be vulnerable to malfunction and manipulation and can easily be made to generate spam and malware. 

• Environment. OpenAI has not revealed the environmental impacts of its products, including ChatGPT.

• Education. ChatGPT has fueled concerns that students can easily generate high-quality essays and other material, thereby reducing their need to build critical-thinking and problem-solving skills.

• Politics. ChatGPT's ability to generate plausible-sounding misinformation, disinformation, and hate speech is seen to have potentially serious effects on the well-being of communities and democracy. 

• Employment. ChatGPT is seen to pose a risk to jobs and job conditions across a swathe of industries, from law, finance, and marketing to academics, journalists, and software programmers. OpenAI's development of ChatGPT involves poorly paid workers in third-world countries with limited rights.

Lawsuits ⚖️

• Canada: TORONTO STAR NEWSPAPERS LIMITED, METROLAND MEDIA GROUP LTD., POSTMEDIA NETWORK INC., PNI MARITIMES LP, THE GLOBE AND MAIL INC./PUBLICATIONS GLOBE AND MAIL INC., CANADIAN PRESS ENTERPRISES INC./ENTREPRISES PRESSE CANADIENNE INC., and CANADIAN BROADCASTING CORPORATION/SOCIÉTÉ RADIOCANADA v OPENAI, INC.

• Germany: GEMA v OpenAI [summary]

• India: ANI v OpenAI; Federation of Indian Publishers v OpenAI

• UK: Mumsnet v OpenAI [summary]

• USA: Alter v. OpenAI; Authors Guild v OpenAI; Basbanes v Microsoft Corp; Center for Investigative Reporting v. OpenAI; Chabon et al v. OpenAI, Inc. et al; Daily News v. Microsoft; Doe 1 et al v GitHub; Doe 3 et al v GitHub et al; Mark Walters v. OpenAI; P.M. et al v. OpenAI LP et al; Raw Story Media, Inc. v. OpenAI Inc; S A.T., J.H., vs. OPENAI LP, OPENAI INCORPORATED, OPENAI GP, LLC, OPENAI STARTUP FUND I, LP, OPENAI STARTUP FUND GP I, LLC, OPENAI STARTUP FUND MANAGEMENT LLC, MICROSOFT CORPORATION; Sancton v OpenAI; Silverman et al v. OpenAI, Inc. et al; The Intercept Media, Inc. v. OpenAI, Inc.; Tremblay et al v. OPENAI, INC. et al; Ziff Davis v OpenAI

Investigations, assessments, audits 👁️

• September 2023. Poland investigates ChatGPT for alleged privacy abuse

• July 2023. US FTC investigates ChatGPT for possible consumer harms

• June 2023. Japan warns OpenAI over ChatGPT AI training

• April 2023. Canada investigates ChatGPT privacy concerns

• March 2023. Italy bans ChatGPT over data privacy concerns