In a previous blog, I discussed a general overview of Artificial Intelligence/Machine Learning (AI/ML). If you missed it, click here to read it. This time, let’s get into the weeds of AI and talk about its use in software development.
AI is no Overnight Sensation (Even if it May Seem Like It)
AI summaries are popping up on our social media comments and at the top of search engines; AI-created images are spreading rapidly across the internet; AI bots have even saturated social media platforms. For those who don’t have their finger on the artificial pulse, it seems like AI has become a sensation overnight.
However, this futuristic reality has been building for decades - three quarters of a century to be exact. Many people don’t know AI began in the 1950s or that it goes so much deeper than the aforementioned examples.
Generative AI - We All Know It
By now, one of the most popular forms of AI is Generative AI. Think: OpenAI’s ChatGPT or some equivalent. ChatGPT is a chatbot that spits out responses to a user’s prompt in seconds. Ask it to write a poem, change the tone of your email, write a shopping list, or answer a question and it will do so instantaneously. Incredible! How does it do that?
These chatbots use Natural Language Processing (NLP) and Large Language Models (LLMs) to learn tons of data, recognize statistical patterns (and natural human speech patterns) and use deep learning, like Generative Pre-Trained Transformer (GPT), to process and generate human-like text. To learn more about NLP and LLMs, check out Trinity IT’s blog covering them.
Now we know what the GPT in ChatGPT means. It’s time to get into the weeds.
GPT in Software Development
AI is transforming software development by handling the time-consuming coding tasks for developers, so they can focus on the bigger picture: business plans, creative creation, architectural planning, etc.
According to a Forbes article titled “AI Is Driving An Evolution In The Role Of The Software Developer” by Gary Drenik, software developers who use AI to code could code 126% more projects each week.
Using AI for repetitive, tedious tasks is the key to higher productivity and more brainpower for creativity and strategic planning. Plus, it’s making software development more accessible and expanding the developer community.
“Using [Application programming interfaces (APIs)], AI not only democratizes business analytics integration for developers who don't have specialized data knowledge but also advocates for a more integrated approach where analytics are built as fundamental components of applications. This shift from traditional business intelligence (BI) to an embedded analytics approach allows developers to create rich data experiences without in-depth knowledge of BI, SQL, or data modeling, thus opening analytics to a wider developer community,” writes Drenik.
Combining AI and software developers’ own expertise results in more innovative products in a fraction of the time. The current GPT models are so advanced that they can build websites and write code, making them an invaluable tool in this new era of software development. AI is celebrated as the thing that boosts human productivity by saving time on tasks that would otherwise take hours.
Stay Secure
Next month is cybersecurity month, so it’s only right that we emphasize the importance of keeping your data secure when working with AI. As a government contractor, Trinity IT is especially concerned with maintaining the highest security measures with all of our data, personnel and equipment.
When using OpenAI’s ChatGPT, exercise caution. ChatGPT is NOT certified to handle classified or highly sensitive data under any government or industry-specific regulatory standards (e.g. Federal Information Security Modernization Act (FISMA), Health Insurance Portability and Accountability Act (HIPAA), General Data Protection Regulation (GDPR)).
For handling classified or sensitive data, always use a secure, approved system designed for that purpose, rather than public AI services like ChatGPT.
OpenAI offers business and enterprise solutions where data control and privacy are stricter. If you are considering using AI models for secure or confidential purposes, it’s better to explore those enterprise offerings or work within internal frameworks that comply with your organization's security policies.
Like Humans, AI has its Flaws.
No one is perfect, not even a robot (at least not yet). For now, we humans still must have a foundational knowledge of, or ability to research, the topics we are asking our favorite AI bot about. Time and time again, AI has been proven wrong, sometimes on very simple questions. So, it’s important to remind ourselves that we cannot rely on AI to entirely replace our common sense, our research skills or our creativity.
It’s easy to ask AI to write a blog for you, but it takes away the personality in the writing. It’s a quick solution to enter a prompt and have AI generate a cartoon, but it takes the joy and creativity out of creating art for yourself. Want to learn about home cool historical figures? Sure you can ask ChatGPT, but make sure to supplement with your own research to ensure you’re learning the most accurate information.
AI is incredible and should be recognized for its achievements. It’s a huge sector of the tech world that is taking the spotlight, but don’t forget your capabilities. It’s supposed to help us - not replace us.
Elizabeth Casalnova,
(267) 396-7901
elizabeth.casalnova@trinityit.biz