Dr. Norman Packard is trying to create life in his lab. This doesn't involve corpses and bolts of lightning... or at least, not yet. It does involve lipid chemistry, genetic algorithms and lots of questions about the nature of life.
Packard defines a living system as one that's capable of self-maintenance, self-reproduction and evoleability. This definition is pretty broad - does flame qualify? Turbulence, plasma, stars? Packard believes evolvability is the tough part - it's more than change - it's inheritance and selection. While he thinks flame doesn't qualify, he thinks societies, ecosystems and the Internet might...
Even extremely simple lifeforms, like the E. Coli bacteria, are significantly more complex than we can currently engineer. If you keep them fed, they reproduce - that's more than our most complex technology ca currently do. We don't know how to engineer a system that serves as a factory as well as the product.
This is the problem that Packard and colleagues - working on a project called PACE - Programmable Artificial Cell Evolution - are tackling. They'd like to build systems that have a container, a metabolism to harvest energy, and genetic information, either stored in DNA or PNA (Peptide Nucleic Acid, which doesn't exist in nature...)
Packard's group works on the container, using vesciles made of lipids to contain the other "cell" structures. There's an enormous range of possible structures - shapes - for these cells. Packard is using genetic algorithms with fitness functions around size, growth, divisibility and the possibility to carry a payload to decide what type of lipid and what shape his cells will be using.
In this sense, Packard mentions that he's doing "postmodern science". In modern science, as he defines it, one documents natural laws and can derive results from those laws - in postmodern science, like evolution, there are emergent behaviors that can't be derived from a starting point. Packard believes that we need to accept limitations on derivability, give up on engineering every detail and learn to live with the uncertainty that results.
Why create artificial cells? Well, they could be hugely useful, for smart drug delivery, diagnostic tools, environmental remediation, energy or carbon sequestration. On the other hand, they could be terrifyingly dangerous, as a number of questioners online and in the audience pointed out. What if someone starts programming cells to carry terrible diseases? Packard believes that this is the problem anyone faces with a sufficiently powerful toolkit... and also believes that the development of this powerful toolkit is, basically inevitable.
There is something to remember here. It will be hard for these artificial organisms to thrive in the wild unless we intentionally design these abilities into them.
For example, these artificial microbes could require a few exotic molecules as food, molecules not found in nature. In the wild, they'd starve. If Packard's critters use PNA, it will be very hard for them to interact with DNA-based life. It is unlikely they'd thrive outside the lab. As long as we design each new organism so it can't live outside the lab, the risk is reduced.
Unlike GMO, the real danger here is abuse, not merely unintended consquences or accidents. There could be bio-spammers that build bacteria that write advertising graffiti on your walls. Science fictional I know but, never say never.