The attention of the technology world has turned to artificial intelligence. Companies across sectors are asking how they can become more deeply involved in this technology, developing it, using it and investing in it.
The excitement has spread beyond the tech sphere, with the name "ChatGPT" becoming a household word. One reason why AI has caught on in such an overwhelming fashion is its wide variety of use cases. It's easy to imagine common types of AI, such as generative language models, assisting employees across industries and roles.
In light of this widespread potential, it's worth taking a look at the state of AI, which considers several different parallel paths. AI is many things at once — a revolutionary tool for simplifying repetitive tasks, an object of tech industry hype and an uncharted legal area.
When looking at the current state of AI, as decades of development harden into a new status quo, a few of the key topics to review include:
Though these three concepts don't represent AI's full potential, they help illuminate several of the key capabilities and limitations of the technology, while offering hints about where it's going. Participants in sectors that could soon see increased AI presence, such as software engineering, can better understand their own businesses by considering what's next for these algorithms.
Internet search is one of the areas where developers see massive potential for AI to change how the whole field operates. The current process of typing in search terms and scrolling through possible results has been in place for years, but it's natural to wonder: Could an AI algorithm send a searcher to the most relevant result for their query, no results page necessary?
Generative AI algorithms are trending toward a simplified search experience. These programs take their raw information from large data sets — such as the contents of discussion threads on Reddit, and aggregate that content to make recommendations that align with searchers' interests and values.
That ability to calculate a match based on variables such as interests and values is where AI truly stands out. By paging through enormous stores of data to find out in-depth details about a business, an AI algorithm can recommend that a user work with a company they've never heard of before, based on shared values such as environmentalism or political advocacy.
Recommending things that searchers don't realize they need yet represents one possible way forward for AI.
Rather than acting like present-day search engines, AI can play an "executive assistant" role, fueling quick decision-making by giving informed advice.
Typing searches into Google is an extremely common action. That being the case, it's worth asking how an infusion of AI technology will change not just the process of receiving search results, but also the way people act. Access to quick, decision-making insights is bound to have an impact on behavior.
One likely outcome is a combination of increased efficiency and decreased capability. This means that people will complete most of their search activity more quickly than ever before, but will lose their ability to make decisions without the aid of AI assistants.
For a quick example of this phenomenon in action, it pays to look back at the early days of GPS in cars. The devices allowed drivers to get to their destinations with no maps — but as they came to rely on GPS assistance for even short trips, people on the whole became less able to navigate their towns by memory or to read analog maps.
The AI-driven search era may make some common searches drastically easier while lessening users' abilities to rely on their intuition or pick the most relevant search result from a list.
Software engineering with the assistance of AI tools seems set to change, at least in terms of day-to-day processes. It's important to remember, however, that the fundamental concept behind coding — creating a product to solve a problem — is relevant no matter what technology engineers have access to.
As generative AI models become more powerful, these tech tools will likely have a role in the coding space. Since AI can't actually think, the work performed by algorithms will probably take the form of brute-force coding of standardized features.
With AI churning out large blocks of usable programmatic code and sifting through code in search of bugs, that leaves human software engineers with more time in their days. These free moments can be put toward solving novel problems and developing industry-specific concepts, valuable activities that can't be easily shifted to AI.
One interesting area to watch in software engineering is the creation of user interface and user experience elements. With enough training, it's feasible that AI algorithms will be able to produce fully functional, if unspectacular UI and UX. In such a case, human development would become rarer and largely apply to high-end or specialized app experiences.
One open question about AI and software engineering has to do with the future of current development tools such as the Flutter multi-platform framework or the GitHub repository. These technology tools and others like them are owned by tech giants such as Google and Microsoft.
If those large companies decide that AI is the primary driver of their business going forward, they could push developers toward AI-assisted coding by ending support for the current tools. This would mirror the kind of thinking that caused Google to end support for its RSS reader, shifting the way web users interacted with blog content.
Whatever form AI-powered software engineering takes, there will be a role for human engineers, taking on the decision-making and creative problem-solving jobs. Furthermore, there are a few issues preventing companies from fully investing in AI as a development tool — such as unclear intellectual property rules.
What is the IP status of an idea, a concept, an image or a piece of code created by a generative AI algorithm? The answer to this question is currently unsettled, putting full-throated corporate endorsement of AI in a state of limbo.
Perhaps the best historical precedent for AI's current status comes from the moment when open source development first became popular. In those days, up to roughly 2010, companies adopted policies forbidding their teams from using open-source code, even when it could solve problems for them, due to the unclear legal picture around control and ownership.
In the months and years ahead, it will be up to the courts in various jurisdictions to determine what defines a "derivative" work. Once these matters are settled, it will be easier for stakeholders to draw up policies around the use of AI, and from there, the technology will have an easier time taking hold in software engineering and beyond.
The complex legal picture around AI partially comes from how these algorithms work and the way that functionality interacts with privacy laws. AI tools draw their data sets from vast amounts of content generated by internet users — so what happens when those individuals claim control over their information?
The European Union's General Data Protection Regulation (EU GDPR) allows individuals to insist that their data not be used for specific purposes. How would such a request work if that information had been used to train an AI algorithm? Would the owner of that system have to retrain it without the data? This would limit the efficiency of AI's ability to process a data set once and use that information for years to come, thus changing the value proposition of AI.
Unresolved legal issues such as the question of GDPR compatibility, are a significant part of the AI story today, and the answers will shape its immediate future.
AI, despite its decades of development and surge in popularity, is still very much a technology in transition. At this moment, it is bringing a general change of emphasis in the way companies think about projects and processes. Tasks that algorithms can't take over — creative thinking, strategic development, domain-specific expertise — are receiving new emphasis.
Future pipelines and workflows will be conceived in a world where AI exists. This will affect how these business processes are designed, with repetitive or non-creative tasks increasingly eyed as places for AI to take over.
To start thinking about the future of design and development in this rapidly changing climate, it can pay to work with experts who are at the cutting edge of their industry. This is where a consulting engagement with Transcenda can pay off, infusing your team with strategic expertise when you need it most.
Contact Transcenda to discuss the possibilities for your organization.
