Generative AI holds both promise and risk for global brands, let alone college students seeking help with their homework. Simply defined, generative AI creates different types of content such as text, images, and audio based on machine learning language models that can consume large amounts of text. This content may appear to have been created by humans.
Generative AI has become more accessible thanks to new tools like ChatGPT, and the latest version promises to transform many business sectors. Meanwhile, regulatory uncertainty hangs over AI adoption as governments look to “accountability mechanisms” for technology.
Dr. Mukesh Dalal, an AI and machine learning expert who builds business applications using generative AI, will join a conference call hosted by GLG’s Abhishek Tyagi to discuss the business and regulatory implications of generative AI. , discussed his future predictions. Here’s an edited summary of the discussion:
Could you share your insights on emerging AI business models?
AI business models are evolving rapidly. Computational costs are reduced while more data is available to AI. This has attracted the attention of many companies. AI brands and vendors that are not typical AI players are ramping up their products with AI capabilities. Hardware brands such as NVIDIA and Qualcomm are developing hardware for AI use cases. Public cloud brands such as Amazon and Microsoft are adding AI to their current services and creating new AI-based services. Database brands such as Microsoft and Snowflake have enhanced their data capabilities to support use cases.
New brands are also developing AI-focused products. Databricks creates data services for AI, and Snorkel allows people to label AI training sets. Companies like OpenAI are creating general-purpose pre-trained models like his GPT, and Siemens is creating industry-specific models.
What are the intellectual property (IP) and plagiarism risks in generative AI technology?
One of the biggest risks is that music, videos, or text articles can closely resemble copyrighted material. That’s a clear violation of copyright law. Another area is where students are using generative AI systems to complete their homework.
Are there any new solutions to solve this and what are the challenges of Watermark AI?
Watermarks can detect if the content was generated by AI. This is a difficult problem and most people believe it is unsolvable. It is impossible to safely watermark all AI-generated content. Technology evolves, and so do the people who abuse it. It’s a cat-and-mouse game with no perfect solution.
Another solution is information disclosure. When you use Generated AI Content, you are disclosing it. It may work if this becomes mandatory, but there will always be people who violate this and go undetected.
Can you tell us a few things about current US AI regulations?
The federal government is reluctant to impose regulations that could slow AI innovation or put the U.S. in a weaker position than China and others. AI regulation in the US is led by states. California is a leader in generative AI regulation. Certain features, such as facial recognition, are regulated and prohibited in some jurisdictions within the United States.
Another example is ChatGPT, which has been banned in New York City schools, but it’s unclear how long that will last, as some educators claim it can be a powerful tool. Several universities actually encourage his use of ChatGPT.
How would you rate the EU Artificial Intelligence Law?
The European Union is creating regulations that apply to all EU member states. The law was adopted in December 2022, but countries such as Germany are calling for improvements. Not sure how soon this will be implemented. Her two European standards bodies are helping to clearly define and apply AI law.
Other applications will be heavily regulated. Suppose a company uses AI to scan job candidates. AI can introduce bias and put certain applicants at a disadvantage.
Do you think AI methods address the challenges that arise across generative AI segments such as images, text, and artwork?
The development of generative AI is accelerating beyond how AI methods are designed. ChatGPT did not expect it. Some risks are not covered by AI law, but may be incorporated when standards bodies draft details. It will need to evolve to deal with various risks.
What regulations will be required when comparing different sectors and applications?
This is a controversial topic. Expert opinion is divided. There are some existing regulations around product safety and defamation that should also apply to AI. Ideally, existing regulations would apply to AI first.
In some cases, new adverse effects have arisen based on current regulations. For example, Section 230 of the US Communications Act protects Internet platforms from being treated as issuers. If I post anything defamatory about anyone on Facebook, Facebook is not the publisher and I am legally responsible. it needs to be updated.
How do you see the impact of AI technology on the audit market?
AI is clearly both an enabler and a risk for businesses. Risks increase as companies adopt AI. Understanding how much risk your business is exposed to and how to mitigate it is critical. This suggests that there may be growth not only in AI, but also in the AI audit industry. It’s also for data. Audit scope includes data used to drive AI applications.
AI will join accounting auditors, who are already audit experts. This could make him one of the fastest growing areas of their business. Deloitte and PricewaterhouseCoopers are already advanced in data and AI with their consulting businesses and could be leaders. Other auditors such as EY and KPMG will also evolve their businesses to focus on AI and data auditing.
What are your thoughts on generative AI changing the media and entertainment industry?
Generative AI will wreak tremendous upheaval. Media and entertainment companies rely on generating new content such as new movies, scripts and songs. Many professionals have traditionally created such content, and the industry has been paid heavily. Non-professionals used to create content, but the quality was low. My home video posted on YouTube is nothing like a Spielberg movie.
With generative AI, quality can improve rapidly and become difficult to distinguish from professional content. AI tools will cost less and the differentiator could be marketing and promotion. Other results are possible. Fewer professionals and artists may be employed, but the wages of those employed will be higher. Massive cataclysms are expected in the next few years.
What are the effective solutions to the threat this technology poses to artists?
Artists will need to find ways to maintain their reputation and work. They may focus on different types of content, promote that content, and increase production. Even if the cost per unit goes down, you can make up for it in quantity. If each artist generates 10 content pieces today, ie 10 videos of her, it could generate thousands of content her assets per year.
What is your outlook on emerging players?
I would expect cloud providers to continue adding AI-based revenue-generating services or continue adding AI to existing services to generate revenue. Companies such as Amazon, Google, and Microsoft will benefit from AI adoption. They will deploy it in the public cloud and get short-term benefits. Other organizations with short-term advantages are companies like OpenAI, which provide tools to create and deploy AI capabilities.
Data companies will likely use AI and data clouds, competing with current public cloud vendors. These models have uncertainties. They will evolve, but it is unclear who will win.
It’s going to be a wild ride. More regulation, more risk. In the short term, you will not know how to deal with risk. When you look back in five years, you’ll be amazed at the growth that AI has caused.
About Dr. Mukesh Dalal
Dr. Mukesh Dalal has expertise in AI and Machine Learning and is currently the Founder and CEO of AIDAA, focused on building business applications with GPT-3 and DALL-E-2. Previously, he served as Chief AI Officer at Stanley Black & Decker Inc., leading the analytics and automation strategy. He also served as Global Head of AI and Data and Chief Analytics Officer at Bose Corporation. Prior to that, he was Principal Scientist at Charles River Analytics Inc. and Principal Scientist at BAE Systems where he led research and development in data science and machine learning. Mukesh was the Chief Executive Officer of Big Data, Chief of Webtrends, where he was also an architect and scientist. Mukesh has a PhD in Artificial Intelligence.
This tech industry article is adapted from the GLG conference call Regulating Generative AI: Legislation or Industry Self-Regulation? If you would like access to the recording of this event or would like to speak with technology industry professionals such as Mukesh Dalal and our nearly one million industry professionals, please contact us.
Questions posed during the conference call:
- Could you share your insights on emerging AI business models?
- How do you see the medium- to long-term sustainability of these business models?
- What are the risks of intellectual property and plagiarism? Do you see this as a threat to generative AI technology as a whole?
- Are there any new solutions that can help solve these problems? What are the challenges of watermark AI solutions?
- Can you talk about some of the current regulations on AI in the US?
- How would you rate the EU Artificial Intelligence Law?
- Can EU AI laws address the challenges that arise across all generative AI segments (images, text, artwork, etc.)?
- What are the pros and cons of different approaches by different government agencies?
- What kind of regulation is needed, legislative action or industry self-regulation, comparing different sectors and applications?
- What impact will artificial intelligence technology have on the AI audit market?
- Looking at the AI Audit market status and key players, what are the competitive advantages?
- What are your thoughts on the changes in the media and entertainment industry observed by generative AI?
- Given the rise of generative image, synthesis, and audio technology, what are the effective solutions to the threat this technology poses to artists?
- Are your companies ready to successfully deploy generative AI against the backdrop of the associated risks?
- What are the short- to medium-term prospects for the sustainability of emerging players and business models considering industry challenges?
