We asked experts what is the most significant realization that business and society will need to address regarding Gen AI in 2023, and how will Gen AI change the internet for the average user?
"While everyone has a mental model for generative AI uses, in 2023, they’ll realize that generative AI touches myriad domains, from education, to politics, to public health, and culture. Crafting policies that ensure the technology positively impacts society is critical. In 2023 geographically diverse stakeholders from many sectors will need to work together with technology companies to support generative AI that furthers human endeavor. In particular, 2023 will be the year to experiment with tools for disclosing when content has been generated with AI and ensuring disclosure tactics are resilient to manipulation and/or forgery."
– Claire Leibowicz, Head of AI and Media Integrity at Partnership on AI
"2023 will see generative AI deeply embedded into the tools, services, and media that ground our digital lives. Businesses will need to consider not just how they could adopt and benefit from generative AI, but also how doing so could generate new ethical obligations to their customers and society. As the distinction between authentic and synthetic or real and virtual becomes increasingly tenuous, transparency about how a piece of media is created will be essential to avoiding further corrosion of our already fragile information ecosystem."
Featured Content of the Month: Finland-based artist, Antti Karppinen, recreates his own images using AI.
Headlines to Note
Gen AI and Education: Pressure on mainstream platforms to identify and disclose generative AI outputs will increase. Public schools in New York City have restricted access to ChatGPT on school networks and devices due to concerns over the negative impacts on student learning and the safety and accuracy of the content that ChatGPT produces. OpenAI said it is developing solutions for spotting text generated by ChatGPT. Also, a student at Princeton, built an app to detect text written by ChatGPT, highlighting the growing need for transparency in synthetics.
Threats: Synthetics will worsen phishing and other cybercrimes exponentially. Cybercriminals are already using deepfake technology to carry out sophisticated attacks. A recent VMware survey stated that 66% of respondents had witnessed deepfakes used for attacks, and lD fraud.
Courtroom Problems!!: The existence of deepfakes and the widespread adoption of generative AI will impact the use of video evidence in legal procedures. Wilmerhale’s Matthew Ferraro and Brent Gurney recently explained that synthetic media increases the risk of falsified evidence and makes it more likely for parties to challenge the integrity of evidence.
Courtroom Innovation??: Generative AI may also be innovating courtroom appearances by equipping defendants with chatbot technology in their ears! Politico reported that Do Not Pay, an AI-powered legal service dubbed “the first robot lawyer,” plans to “send a real defendant into a real court armed with a recording device and a set of earbuds.” Those earbuds will be delivered AI-powered legal advice to an unknowing judge.
Copyright: Experts predict that Generative AI will lead to copyright issuesin 2023. Digital creators argue that some Gen AI programs infringe on the rights of artists and lawsuits are already beginning. On Jan 14, artists filed a class action lawsuit against Stability ai, Midjourney, and Deviant Art alleging that those Gen AI companies are infringing on the rights of artists and other creative individuals under the guise of alleged artificial intelligence. As Lensa and other similar apps increase in popularity, the pressure for legislationwill grow. Some creators are moving towards opting out of data setsas legal challenges lie ahead.
Next Gen AI Use: 3D model generators will be the next big trend in the Generative AI world. OpenAI released a machine learning system called Point-E that creates 3D Models Using AI and image-to-3D generation. Experts believe that text-to-3D synthesis will likely inspire more advanced models soon.
More Advanced: The next generation of OpenAI’s generative language model GPT-4 will be released early in the new year. GPT-4 performance will improve in its ability to retain and refer back to information and in its ability to summarize a large body of text.
Correction: In the last issue of this newsletter on Dec 14, it was incorrectly stated that Open AI created DALL-E mini. DALL-E mini is not affiliated with OpenAI.
Our Recommendations
An indispensable guide to understanding how we experience the world within and around us and find our place in the universe by Antonio Damasio: Feeling & Knowing
Synthetic media is seeping into our everyday lives, but are we ready? Henry Ajder examines the legal, political and ethical implications - starting with deepfakes’ murky origins: The Future Will Be Synthesised
Have any comments, ideas, or opinions - send them to us:
