Be aware of AI data collection

With the large-scale rollout of generative AI and large language models, tech companies are scrambling for more human-generated text to feed into their training data. Social media platforms, news outlets, and publishers are selling their content libraries to companies that want data for training these models. However, the way this data is collected and sold is often far from transparent.

For instance, on 19 September, LinkedIn - owned by Microsoft, which also has a significant stake in ChatGPT's parent company, OpenAI - quietly opted all users into a setting that permits the use of personal data for AI content creation. 

This potentially included everything from profile information, posts, articles and other contributions to potentially even private direct messages. In response to privacy concerns, the next day LinkedIn temporarily suspended the use of UK user data for AI training after the UK's Information Commissioner's Office (ICO) raised concerns.

This tactic is part of a broader pattern where companies revise their terms and conditions to allow extensive use of personal data - public or private - for training AI models. Often, these changes occur quietly, with users unknowingly defaulting to new settings that permit such use. The lack of transparency and express consent is, of course, unethical but it happens regularly.   

For psychologists, this issue is particularly important. Sharing personal information, client data, or even casual comments on platforms like LinkedIn could result in that data being used to train AI models, with no control over how it might later be reproduced or applied. For occupational psychologists, it raises serious ethical questions about data privacy, consent, and the potential misuse of sensitive information shared in professional settings.

Social media platforms are increasingly turning our content into AI training data. Be mindful of what you share, never share confidential data using social media apps and keep an eye on your privacy settings.

But it's not all bad. Generative AI is moving quickly and there are opportunities to use AI effectively and ethically alongside work psychology. Read more in this recent BPS blog about generative AI, and some of the current practical applications and ethical considerations.

Report by Ian MacRae, a member of the BPS's Division of Occupational Psychology committee.