Syed Imran Haider 
The European Union has rolled out a new set of rules aimed at monitoring the use of artificial intelligence (AI). As per these laws, companies would have to be more transparent about how they use data for training their systems. This is considered to be one of the most heavily guarded secrets in the industry.
The rise
It was 18 months earlier that Open AI, which is backed by Microsoft, launched ChatGPT to the public. Since then, generative AI has seen a massive surge in investment and public engagement.
Generative AI refers to applications capable of producing audio content, images and texts quickly. The industry is expanding very rapidly.
Therefore, it has given rise to questions about how AI firms obtain the data they are using for training their generative AI models.
There have also been questions about copyrights, as these models are fed Hollywood movies and bestselling books without the permission of creators.
To address these issues, the European Union recently passed its AI Act. This will be put into effect in the course of two years.
It would provide regulators with enough time to implement the new laws and businesses would have to deal with new obligations. However, it is unknown how some of the rules would be implemented.
The complications
One section of the Act that has been deemed contentious requires companies to provide ‘detailed summaries’ of the content they are using for training their generative AI models like ChatGPT.
The AI Office, which was established recently, said that it would provide organizations with a template they can follow.
The said template would become available in 2025, after they have a consultation with stakeholders. The details may not have been clarified as yet, but AI companies are certainly unhappy.
They are not willing to disclose what they have used for training their models. It is regarded as a trade secret that would give an unfair advantage to rivals, if made public.
There will be big implications on both small AI startups as well as big companies like Meta and Google, as AI has become the center of their future operations.
Trade secrets
Creators have filed lawsuits against a number of prominent tech firms in the last year, including OpenAI, Google and Stability AI.
According to these creators, these companies used their content improperly for training their models. A number of executive orders have also been passed by US President Joe Biden.
These ordered were to address the security risks associated with AI. But, there is still no clarification over copyright issues.
There has also been support in Congress to ask tech companies to pay creators for their content. This has prompted these companies to sign content-licensing deals with websites and media outlets.
For instance, OpenAI signed a deal with The Atlantic and the Financial Times, while Google made a deal with Reddit and NewsCorp.
However, OpenAI has faced a lot of criticism and backlash for its practices. This included action from actress Scarlett Johansson for using her voice in the newest version of ChatGPT.
