A Glimpse into the Future: How AI Will Impact Our Jobs in the Next 3 Years

Whatever AI can do well, most companies won’t need humans to do the exact same work. That’s a fact. But like with every new transformative technology, the impact of AI will be that some old jobs won’t be needed anymore, some will be redefined and a lot of new ones will be created. If you are looking for a detailed summary of what AI can do now with easy-to-replicate examples and what it potentially can do in the next 2–3 years, then this article is for you.

Generative AI, Large Language Models and their specialized siblings, the Fine Tuned Models, allow us to perform feats in a fraction of the time they previously took, be 10X versions of ourselves as I like to term it. In today’s digital landscape, AI’s capabilities to generate data extend beyond traditional boundaries. From text, image, audio, and video to software, graphs, charts, and tabular data — AI is flexing its muscles in novel and promising ways.

And we are still scratching the surface here, the potential is much much more. Every profession that relies on using pre-existing knowledge to output some form of content, is poised for a significant upheaval.

Co-Working with AI (courtesy — Midjourney)

Generative AI: The jobs it can do right now

When I say that Generative AI can do this and that, I don’t necessarily mean ‘on its own’. It does need human guidance and oversight. But, the humans who can guide and oversee AI to produce the desired output are now 10 times faster than the existing ‘experts’ who are doing the same job now by leveraging their ‘knowledge and expertise’.

Text Generation and LLMs — ChatGPT is literally, the ‘knowledge of the world’ at your fingertips and expert advice a click away. From students to content writers to professionals in every industry, it is making lives easier for everyone. I won’t spend any time here on the specifics, since if you are reading this, chances are that you already use ChatGPT or at least are aware of how others are using it.

Image Generation — For social media marketing, arts, gaming, and the retail industry, the possibility of AI-generated images are already reshaping the landscape. Whether it’s designing unique logos and art, creating personalized items for gifting or graphic design for social media, the list is endless. Midjourney, Leonardo.ai and StableDiffusion are already at the heart of these industries at present.

Audio and Video Generation — This has profound implications for short form videos, animation, games, and even movies. Meta’s voicebox technology has already been showcased to synthesize authentic-sounding voice recordings from a written script, given just three minutes of audio from any individual.

Softwares like Gen-2 from Runway, Kaiber, WonderStudio and so many others are groundbreaking to say the least, reducing the time and skill required for video creation, editing and animation by a huge margin.

Software Development — Combined with low code technology, software creation has become exponentially simpler, making complex programming accessible to a wider audience. Case in point, I developed a POC for a sophisticated bot that automated the homework distribution process for students, auto-checks the responses and provides a 24*7 mentor to students which can also impersonate famous personalities, improving engagement rates by 5700% and reducing mentor effort to near zero. All this took just a couple of days, a task that would’ve required extensive programming efforts just three years ago. Use any low code automation platform like botpress.com to get started or use GitHub CoPilot and ChatGPT to work with Python, Langchain and Open AI APIs to quickly create something which someone in 2017 would term as miraculous!

Data Analysis — Data Analysis has and can become exponentially simpler with AI being able to write SQL commands from text prompts. AI can analyze and generate reports, graphs, charts, revealing trends and patterns with minimal human intervention. This means anyone, regardless of their SQL knowledge, can analyze and make sense of complex data.

Easy to Replicate Examples you can try now

Lets look at some examples of how we can stitch together some AI tools to make work easier:

Marketing — Anyone without the skills / knowledge for content writing, graphic design

Blog and post writing — Potential workflow at this point in time (Time Required — 30 mins)

1. Figure out a topic for your blog post (use something like ChatGPT, copy.ai or jasper for that as well if you are feeling really lazy!)
2. Search for your topic related and popular keywords on a tool like Keyword Planner, SEMrush, Moz or automate that as well via an AI powered tool like Zupyak
3. Use ChatGPT, copy.ai, jasper or anything similar to brainstorm the sections, sub sections and content of your blog
4. Use Perplexity.ai for cases where you need the latest data from the internet. Voila! You have an SEO optimized blog in 30 mins!

Image Content Generation — This can be used for generating images for your blogs, posts, memes, social media thumbnails, YouTube video thumbnails, 2D and 3D Game Assets etc. (Time Required — 5 to 10 mins)

1. Optional step — Use ChatGPT prompt perfect plugin or any prompt database app like the one on HuggingFace to generate the perfect prompt for a particular tool
2. Use a tool such as Midjourney or Leonardo.ai to generate the image. Iterate till you get exactly what you want and edit it on Photoshop if you have to. That’s it!

Video Creation — Social Media marketing, YouTube videos, YouTube & Instagram shorts, Game Assets. Here is how you can create your custom AI avatar and have it say anything you like with proper facial movement for an Instagram/YouTube short video (Time Required — 30 mins)

1. Use Leonardo.ai or Midjourney to create your character
2. Use ChatGPT to come up with a short captivating text script (e.g. 5 things that a startup founder must know)
3. Create the voiceover using Eleven Labs, murf.ai or dubdub
4. Join these 3 things together to create a video of your avatar talking through the script using did.com or any other similar tool
5. Use Capcut or any other free video editing tool to put captions on your video. Done!

Software Development — From live assistance during coding to generating apps/features in minutes, a lot is possible here. I have provided 2 examples apart from the example given earlier here so as to provide a better understanding.

1. Send personalized messages and email for cold outreach — One of 1000s of examples of automation of repetitive tasks (end to end Time Required — 2 hrs, Knowledge Required — Zapier with a bit of javascript) — Create a Zapier flow to take Lead Names, any relevant information and their LinkedIn/Twitter Profiles, then scrape the profiles, draft a personalized message/email, submit them for human approval or send them directly.
2. Design, create and publish a website (Time required — 30 minutes of experimentation, no graphic design tool skills required, Figma Knowledge along with a limited understanding of web development) — Use Midjourney or leonardo.ai, give it the reference of any existing website or dribble design you like and ask the AI to come up with a website or app design with features/tabs/specifics you want.. And that’s it, you will have awesome inspirations which you can directly then recreate in Figma. Then use framer, buzzy or locofy to generate the website in minutes from the Figma design. Now its possible to create the starting website or apps for startups at super speed!

What’s Next — What can AI do in the next 2–3 years

AI CoPilot for X — Medicine, Law, Education, Finance and such fields where knowledge from books, articles and paper are mandatory — AI will replace the need for humans for transactional queries, if not for more advanced use cases. Why ask a lawyer or a tax consultant a question when an AI assistant can answer my questions in an instant.

ChatGPT is already the CoPilot for many students, in some cases, much to the dismay of educators. Ahem! Ahem!

In Healthcare for example, it holds tremendous potential for better patient outcomes, gives patients access to accurate guidance at a moment’s notice, all while freeing up time of doctors to cater to patients who need a more nuanced expert attention. The UK government, in recognition of this, has just set a £21M fund for incorporating AI in its National Health Service.

Similarly, we can have CoPilots for Sales, Operations, Customer Support, IT support, Data Analysis and all such functions where the AI can assist humans with the right knowledge or information at the right time — all you need to do is to ask properly !

A new era for Internships — With generative AI at the helm, a new world is taking shape where AI is the best intern, capable of working across different domains — from marketing, branding, finance, tech, procurement to sales. Companies will want that 1 intern who can effectively use AI to do all these as compared to hiring different interns for each function.

Videos, Movies and Animations — AI is already and will reduce the skill barrier for video editing, creation, animation and VFX by mammoth proportions.

With AI-powered content creation tools, if you have the creativity, know your AI tools and can guide them, you will be able to create images, audio, videos and complex animations at will.

Sure, minor editing might still be required, but 100s of hrs of super-expensive, high skill driven video creation effort will be cut down to minutes. As an example, Indie films will be able to feature great VFX without adding to their costs. Money will stop becoming a barrier to creativity.

Software Development — AI can already write software, albeit basic isolated code and not sophisticated interdependent ones. But thats next, and when software is able to write software, then why need programmers. At the very least, you would only need experts who are above the level of expertise which AI can achieve. So, ironically, software developers will be competing with AI which other software developers would have developed !

Emad Mostaque, CEO of Stability AI, predicts a future where artificial intelligence (AI) will dominate the programming field, stating, “There will be no programmers in five years.” With data revealing that “41% of all code right now is AI-generated,” Mostaque’s vision of a world where software is able to reliably write itself for low to medium complexity apps & features is not far away.

Tasks such as Data Analysis and Reporting will be so easy that anyone can do them and so having dedicated personnel will become pointless. A side effect of this will be that software development and tech skills will no longer have their traditional high barrier to entry — anyone can become a decent software engineer, may not be experts but definitely decent.

There are a hell of a lot of software project managers out there at present. What they mostly do now will be automated, what they are required to do in 3 years will be very different from what they do now.

Games — Games are after all nothing but images, audio, motion graphics and code. Simple yet creative games will need nothing but creativity. Unity is already working on a text to game feature. While that may be in a nascent stage to start with, it will improve radically with time (don’t forget that their developers now have AI assistants to help them churn features out faster). A ‘not so obvious’ advantage will be that businesses will then be able to truly gamify their apps and websites without having to spend 1000s of dollars on it.

Conclusion

The advent of Generative AI is transforming the way we create and consume content. And do not mistake the word content for ‘social media content only’! We need to adapt accordingly and those of us who know how to work with AI in the new world will fare drastically better than the rest of us.