Table of contents
In case you missed it, TheNationalCIOReview published a thought-provoking piece from David Lefkovits on the major phases of AI development, leading to the most recent phases of Transformative AI and the Sudden Hype of explosive progress linked with the growth of Generative AI.
For CIOs and other senior technology professionals at software and tech-enabled businesses, the challenges and opportunities from Gen AI are two-fold:
(1) In general, how to capture the benefits of Generative AI to improve workflows across the business, and
(2) In particular, with respect to the Software Development Lifecycle, how to increase productivity and “flow state” among developers, while managing the legal, security, and compliance risks associated with developers using GenAI tooling such as specific-to-engineering GitHub Co-Pilot and general tools like ChatGPT.
Looking at that second topic in more detail, we see similar recent trends as with AI use overall.
- Adoption is widespread: Recent surveys indicate at least 63% of developers and 97% of Security and DevOps specialists are now using Gen AI tooling
- Benefits are significant: productivity improvements range from 5-50%. Our own research at Sema indicates that improvements of 25-40% are achievable in the near term.
- Risks are clear: to share two examples, developers believe that AI written code is more secure when in fact it is less so; and in the US and Germany companies will not get copyright protection on their code if too much of it is written with GenAI code (German Copyright Act; US Copyright Office, 2023).
David also noted that “The challenge now isn't just technological; it's also educational.” Our team at Sema couldn’t agree more in the GenAI coding context. The biggest challenge that organizations face with GenAI in coding is teaching:
- teaching engineers how to make the most of Gen AI tools while minimizing the risks,
- teaching engineering managers how to be supportive but firm about capturing the value from GenAI,
- and teaching compliance teams how to measure and manage the risks with regulators and customers.
In light of explosive progress, it is our belief as an organization that governance will be the backbone of responsible (and successful) AI implementation. Problem-solving must pace continued evolution. More to come.
Sema is the leader in comprehensive codebase scans with over $1T of enterprise software organizations evaluated to inform our dataset. We are now accepting pre-orders for AI Code Monitor, which translates compliance standards into “traffic light warnings” for CTOs leading fast-paced and highly productive engineering teams. Sign up for the waitlist to get notified when we launch publicly in Q1 2024.
