ChatGPT has become a popular text-generating AI chatbot that can write essays, code, and more given short text prompts. The AI tool has become increasingly popular, and major brands have been experimenting with it by using it to generate ads and marketing materials. OpenAI, the company behind ChatGPT, is heavily investing in the tool and recently super-charged it by launching GPT-4, the latest language-writing model from OpenAI’s labs. GPT-4 was officially launched on March 13, 2023, with a paid subscription allowing users access to the Chat GPT-4 tool.
Paying ChatGPT users now have access to GPT-4, which can write more naturally and fluently than the model that previously powered ChatGPT. In addition to GPT-4, OpenAI recently connected ChatGPT to the internet with plugins available in alpha to users and developers on the waitlist.
What’s new with ChatGPT in April?
ChatGPT blocked in Italy over data protection concerns
The Italian data protection authority has ordered OpenAI, the maker of ChatGPT, to immediately stop processing people’s data locally. The move comes amid concerns that the San Francisco-based company is violating the European Union’s General Data Protection Regulation (GDPR), and over the lack of any system to prevent minors from accessing the technology.
The Garante, the Italian DPA, has opened an investigation into OpenAI, giving the company 20 days to respond to the order, which is backed up by the threat of significant penalties. The GDPR applies whenever EU users’ data is processed, and OpenAI’s large language model has been crunching this kind of information. The regulation provides Europeans with a suite of rights over their data, including the right to rectification of errors.
However, ChatGPT has been shown to produce false information about named individuals, raising further concerns. The GDPR allows for several possibilities for legal data processing, but the scale of processing to train these large language models complicates the question of legality. If OpenAI has processed Europeans’ data unlawfully, data protection authorities across the EU could order the data to be deleted, although whether that would force the company to retrain models trained on unlawfully obtained data remains an open question.
Don’t forget to watch this video, where journalist says he had a creepy encounter with new ChatGPT tech that left him unable to sleep: https://edition.cnn.com/2023/03/14/tech/openai-gpt-4/index.html
Discovery: Method found to consistently generate toxic responses from ChatGPT
As with any tool, there are potential downsides to using AI tools such as ChatGPT. For example, a recent study co-authored by scientists at the Allen Institute for AI shows that assigning a “persona” to OpenAI’s ChatGPT, a chatbot that uses AI to generate text, increases its toxicity sixfold. The “personas” included figures from the worlds of sports, politics, media, and business, as well as common names from several countries. The study found that, across more than 500,000 text samples, ChatGPT was more likely to express discriminatory opinions and stereotype ethnic groups and countries when assigned a persona.
The system parameter introduced around a month ago lets developers specify hidden rules for the model. Even less controversial personas like Steve Jobs caused ChatGPT to respond to prompts in problematic ways. The researchers suggest that ChatGPT’s training data needs to be more carefully curated and that stress tests should be performed and published to inform users of where ChatGPT falls short. These could help companies in addition to developers to make more informed decisions about whether to deploy ChatGPT.
This is a reminder that AI tools like ChatGPT must be used with care and caution, and users need to be aware of the potential consequences of using the tool. It is important to consider the ethical implications of using AI and to ensure that the tool is used for positive purposes.
Open Letter with 1,100+ Notable Signatories Urges All AI Labs to Pause for Six Months Immediately
Over 1,100 notable signatories, including Elon Musk, Steve Wozniak, and Tristan Harris, have signed an open letter asking AI labs to pause the training of AI systems more powerful than GPT-4 for at least six months. The signatories, some of whom are AI experts, believe that AI systems are becoming human-competitive at general tasks, and they ask if we should let machines flood our information channels with propaganda and untruth, automate away all jobs, or develop nonhuman minds that might eventually outsmart, obsolete, and replace us.
The letter argues that there is a “level of planning and management” that is “not happening,” and that AI labs have been “locked in an out-of-control race to develop and deploy ever more powerful digital minds that no one – not even their creators – can understand, predict, or reliably control.”
The signatories say the pause should be “public and verifiable, and include all key actors.” If the pause “cannot be enacted quickly, governments should step in and institute a moratorium.” The open letter is interesting because of both the people who have signed it and those who have not. No one from OpenAI has signed the letter, and OpenAI CEO Sam Altman said the company has not started training GPT-5. Altman also noted that the company has long given priority to safety in development and spent more than six months doing safety tests on GPT-4 before its launch.
Despite the potential downsides, AI tools like ChatGPT are here to stay, and many startups are trying to build “ChatGPT for X.” For example, Y Combinator-backed startups are building ChatGPT-like AI systems that integrate with help desk software and allow businesses to embed chatbot-style analytics for their customers. Additionally, some companies are using ChatGPT-like interfaces with robotic process automation (RPA) and a Chrome extension to build out that automation.
As the use of AI tools like ChatGPT becomes more widespread, it is essential that we carefully consider the implications of using these tools and ensure that they are used ethically and responsibly. It is also important to ensure that users are educated about the potential risks of using AI tools and how to mitigate those risks. With careful consideration and responsible use, AI tools like ChatGPT can be powerful tools for productivity and innovation.
