Big NOse of?AI

As an engineer, It's not fun to be regulated in cutting-edge technology.

During one of my past project on an AI-driven Call Center optimization, I had to go through 3 rounds of interviews with a privacy officer followed by 1-hour of the test. It was frustrating because I believed that I was doing the right thing and should get a concession in this process, and the privacy officer was putting their nose which was impacting productivity.

However, after gaining a deeper understanding of the data and the regulations, I appreciated the regulations and governance policies. Looking back, I now realize that doing the right thing is not enough; doing the right thing in the right way is what truly matters in the long run. This project taught me the importance of balancing project delivery on time with data privacy, security and other regulations around. It remains one of the best projects I have ever worked on.

Not an one way street:

It is a pressure situation where you must manage the project deliverables and satisfy regulatory compliance. It is even more challenging when the technology practice is new. In that project, I felt Regulatory compliance was like a one-way street, that doesn't consider project success as a priority. For engineers, project success means delivering the expected outcome, and regulatory holdups hinder success. Organizations measure AI engineer efficiency based on outcomes. Engineers and project stakeholders should take AI project regulations as important as project success, else the scale of damage can be bigger than ever anticipated?ref1?ref2. On the other hand, the regulations and governance team need to understand the limitation and provide or help to provide a workaround; only by putting regulations will limit the growth of AI in business which is going to impact business in the near future.

Hidden Cost [3]:

1. Compliance costs:?AI regulations may require companies to invest in new technologies, infrastructure, and personnel to comply with the regulations. This can be a significant financial burden, especially for smaller companies that may not have the resources to implement complex AI systems.
2. Innovation costs:?Regulations can also stifle innovation by making it harder for startups and smaller companies to enter the market. This can limit competition, slow down technological progress, and ultimately reduce the benefits that AI can provide.
3. Competition:?Do you think OpenAI would ever be succeed in ChatGPT if it had regulated its AI development well? The challenge is since there is no standard regulation, someone else will bypass it and go a mile ahead in this development and acquire the market. In addition to that, country-wise regulations are more crucial. For example, AI regulations in Italy could disadvantage companies operating in countries with more stringent regulations.

Regulated AI development:

Developing AI is a shared responsibility. Regulators cannot simply impose new rules and obstacles; instead, they need to collaborate with engineers and stakeholders to mitigate potential risks. When regulators and engineers work together, AI development can be more successful.

Blockages in AI development due to regulations are a failure on everyone's part. The key to success in developing AI lies in working collaboratively with stakeholders, AI governance, and regulators to ensure that ethical and responsible practices are followed.

7 key things to consider in regulations in AI :

1. Source / Owner / Generator of Data
2. Who is the beneficiary of $?
3. Clarity on the project outcome and Use
4. Data security and privacy
5. Speed of innovation under a regulatory environment
6. Metrics to measure Bias and fairness
7. Technical Plan to control/minimize the unforeseen

AI development is overwhelming in regulatory environment. This article is to share my experience and thought in AI projects and the business community to prepare for efficient AI-driven Business with the help of latest tech. I hope this article helpful; feel free to share this and post your thoughts or can reach out to me on balarampanda.ai@gmail.com.

References and Recommended readings:

[1] https://openai.com/blog/march-20-chatgpt-outage

[2] theguardian.com/technology/2023/apr/10/i-didnt-give-permission-do-ais-backers-care-about-data-law-breaches

[3] https://www2.datainnovation.org/2021-more-than-meets-the-ai.pdf

[4] https://infoprotect-archive.mit.edu/risks-to-data

[5] https://futureoflife.org/open-letter/ai-principles/

[6] https://futureoflife.org/open-letter/pause-giant-ai-experiments/

[7] https://www.forbes.com/sites/digital-assets/2023/04/07/the-case-for-artificial-intelligence-regulation-is-surprisingly-weak/?sh=6c784a5b50a8